Digital
Designing complex content in the age of Artificial Intelligence (AI)
October 10, 2025 by deborahamzil No Comments | Category Content Design, Design, Digital, Scottish Government, Social Security Scotland, User-centred design
Guest blog post by Therese Hooper, Lead Content Designer in Social Security Scotland.
This blog is about the tools I used to tackle a complicated content challenge, working in a new cross-functional team facing major time constraints.
The role of content design
You will have heard this before. A content designer’s role is not just writing words. We design the right words, for the right user, at the right time – simple right?
So when tasked with redesigning a content template that can send thousands of government letters every day. I found myself handling ‘the right words’ for users whose ‘right time’ could be anything from 30 different scenarios, each with its own logic, exceptions, and outcomes.
Now the right words, right user, right time takes on a whole new meaning. What happens when content needs to find the right user out of multiple possible user journeys? What happens if everything the user does before reaching the content decides how accurate and sensible the resulting experience is? There was a time where this type of complex content problem would have stretched the content designer beyond limits.
Now, we have generative Artificial Intelligence (AI), structured content, and prompt engineering. Not to replace the content designer role, but to give us the tools and techniques to propel us forward.
Re-imagining the content designer’s toolkit
In an ideal world, content designers and business analysts work on this type of problem together. But in this world, business analysts are hard to come by. Plus, I wanted to understand how this template was going to work. Even in the most complex of cases, my job is to make that end user experience as simple as possible.
I was not the most experienced AI user, but with Microsoft enterprise level protection and a reliable prompt writing framework, I felt confident to give it a go. I shared the problem as I understood it, I need to:
- reach a simple yes or no decision
- explain how results from two independent components led to the decision
- in a specific order, analyse and apply multiple exceptions to the decision
- the outcome should be a draft that can be read by humans and built by a computer
Shaping data and writing patterns
The response I got was more than a list of the ways to deliver the content. I now had a framework for creating a structured approach to meeting multiple cases of user needs.
I refined the prompts, clarified the rules, and built a detailed map of how the template needed to behave. This helped kick start the patterns I needed to write and gave me the kind of detail I needed to understand the different user and business needs for each. I found working this way easier when I had to manage changes along the way and fulfil my part in assuring and confirm the template as it began to move through delivery.
This experience has reshaped how I think about content design. Instead of being handed an Excel sheet and left to interpret it, I collaborated with AI to shape the data into something usable; for Miro, slides, technical documentation. It’s changed how I work.
The future is structured
If you have not yet read Patrick Stafford’s article on the future of content design, go find it. It will inspire any content designer hoping to embrace generative AI. It helped me think of AI as a tool to accomplish more, especially when the work gets technical.
Speaking of technical, I’m also inspired by Rahel Bailie’s work on enterprise-level structured content. Her thinking helped me realise that complexity does not always need to be complicated. By working in components, patterns, and templates, we can set boundaries for good content management and lead better conversations on how we think about content as assets.
You can do this too
This project taught me that AI is not just a shortcut, it’s a collaborator. It helped me turn a dense technical document into a usable logic model. It gave me the level of detailed understanding needed to help test, iterate, and design with confidence. And it helped me build something that works not just for one user, but for many.
So, if you need content to do some complex things for you:
- remember complex does not need to be complicated
- learn some basic prompt engineering and make sure your data is protected
- embrace Patrick Stafford’s future of content design
- read Rahel Bailie
And lastly, please do not be afraid to ask AI for help. Especially when the problem feels too big to tackle alone.
Tags: content design, scottish government, user centred design
Leave a comment